The US president told CNBC on Tuesday that Anthropic was “taking shape” after discussing the Mythos AI model with company CEO Dario Amodei’s chief of staff Susie Wiles and Treasury Secretary Scott Bessent at the White House last Friday.
The Pentagon blacklisted Anthropic A federal appeals court and a San Francisco district court remain in legal limbo, with conflicting conclusions.
President Donald Trump spoke to CNBC Squawk box Tuesday provides a deal anthropicThe AI models to be used by the Ministry of Defense are “feasible” and the company describes them as “improving”.
“They came to the White House a few days ago and we had very good talks with them and I think they’re starting to take shape.” Trump said.
“They are very intelligent and I think they can be very useful.” The comments represent a striking rhetorical reversal from the president, who told Truth Social in late February that he ordered all federal agencies to “IMMEDIATELY stop using Anthropic’s technology” and that his administration would “no longer do business with them.”
Trump’s comments follow Anthropic’s meeting at the White House on Friday, April 18 CEO Dario Amodei met with Chief of Staff Susie Wiles and Treasury Secretary Scott Bessent to discuss the company’s new Mythos model, a frontier artificial intelligence system Anthropic described as highly adept at cybersecurity tasks and has so far only been available to a small group of organizations.
The White House called the conversation “productive and constructive”. Anthropic said Amodei had a “productive discussion” with administration officials about how the company and the US government “can work together on key shared priorities such as cybersecurity, America’s leadership in the AI race, and AI security.”
When reporters asked Trump about the meeting on the runway in Phoenix, he replied, “Who?” he answered. and said he had “no idea” Amodei was there.
The meeting comes amid an unprecedented spat between Washington and the tech industry.
In July 2025, Anthropic signed a $200 million contract with the Pentagon, becoming the first artificial intelligence lab to have its models approved for use on DOD’s classified networks.
But talks stalled as soon as negotiations began in September to place Claude on the department’s GenAI.mil platform. The Pentagon demanded that Anthropic provide unfettered access to its models for all legal purposes.
Anthropic has drawn two firm lines: its artificial intelligence will not be used in fully autonomous weapons systems that select targets without human intervention, and it will not be used for domestic mass surveillance of Americans.
Defense Secretary Pete Hegseth responded in late February 2026 by designating Anthropic a “supply chain risk to national security,” a label previously reserved for companies linked to foreign adversaries.
The official designation, approved by Anthropic management on March 5, required defense contractors to certify that they are not using Anthropic models in work with the military. Trump stepped up the measure with the Truth Social directive.
The designation was unprecedented, as Anthropic argued in subsequent lawsuits: As U.S. District Judge Rita Lynn noted in her 43-page ruling in late March granting Anthropic the preliminary ruling, it was intended to punish public positions of the government for “security,” not an actual national security threat. “Classic illegal First Amendment retaliation,” he wrote.
The legal situation remains divided. A federal appeals court in Washington on April 8 denied Anthropic’s request to temporarily block the supply chain risk designation. Judge Lynn’s preliminary ruling in a separate but related case in San Francisco prevents the rest of the government from applying Trump’s Social Truth ban to Clod.
The practical effect is that Anthropic is excluded from Pentagon contracts, but can continue to work with other government agencies while both cases are pending. DOD continued to use Claude during the US-Iran war, which began before the blacklist took effect.
It is Myth that seems to have changed the position of the White House. CISA, part of the intelligence community and the Cybersecurity and Infrastructure Security Agency, is testing the model.
The White House Office of Management and Budget is developing protocols to allow federal agencies to access the managed version.
Treasury Secretary Bessent’s presence at Friday’s meeting was read by sources close to the negotiations as a sign that the economic and financial security arguments for access to Mythos are reaching the highest levels of the administration.
An administration source told Axios: “It would be grossly irresponsible for the US government to deprive itself of technological breakthroughs. presents a new model. This would be a gift to China.”
Whether there is any recovery Anthropic-Pentagon relations it remains unclear whether it is possible. Trump’s comments on Tuesday refer to negotiations that have promised but failed to produce a deal.
The appellate court’s ruling on the supply chain risk designation still stands. Hegseth did not back down from his position. Anthropic, meanwhile, has engaged Ballard Partners, a lobbying firm where Wills used to work, to advocate around War Department Procurement, showing that it understands the political dynamics as well as the legal ones.
The company has reached $30 billion in annual revenue and is considering an IPO; even if the assignment of supply chain risk does not interfere with commercial transactions, it undermines the credibility of the enterprise.
